[0001] This invention relates generally to an electronic ballast for ignition of a high
intensity discharge lamp, and more particularly to an electronic ballast for controlling
the voltage and/or current applied to the lamp during the initial, non steady state
operating modes.
[0002] In starting a high intensity discharge (HID) lamp, the lamp experiences three phases.
These phases include breakdown, glow discharge, and thermionic arc. Breakdown requires
a high voltage to be applied to the electrode. Following breakdown, the voltage must
be high enough to sustain a glow discharge and heat the electrode to thermionic emission.
Once thermionic emission commences, current must be maintained, in the run-up phase,
until the electrode reaches its steady-state temperature. After achieving the arc
state, the lamp can be operated with a lower level of current in the steady state
operating mode.
[0003] The overall life and efficiency of a lamp are affected by this starting sequence
as are the values and tolerances of the components required to effect this starting
sequence. For ignition, in the pre-breakdown period, the lamp electrode must be brought
to a high voltage for a specified duration. Conventional lamps are characterized by
a minimum voltage level and time duration in achieving breakdown. Typical minimums
range from about 2 to 3 KV for voltage and about 10-100 ms for time duration.
[0004] The high voltage requirements for breakdown can be achieved through pulse resonant
circuits. The frequency at which the circuit achieves resonance and the resultant
resonant voltage varies from circuit to circuit due to variation in component tolerances.
Such variation results in the pulse resonant circuit being designed to withstand nominal
pulse voltages of about 4 to 5 KV, that is, in the circuit being designed to withstand
voltages which are well beyond the 2 to 3 KV range required to start the lamp. An
undesirable increase in cost for the pulse resonant circuit can result,
[0005] Upon achieving breakdown, the lamp enters the non-thermionic glow state. In this
phase, the voltage must be sufficiently high to maintain the flow of electrons. Electrons
are produced by positive ion bombardment of the cathode which produces secondary electron
emission. When the kinetic energy of the positive ions, determined by the cathode
fall, is high enough, sputtering of the electrode occurs. Sputtering of the electrode
produces volatile species of, for example, tungsten which condense on and blacken
the inner surface of the lamp. As the interior of the lamp blackens, transmission
of light through the envelope decreases reducing the visible light level. The pieces
of tungsten which are deposited on the wall absorb radiation thereby heating the lamp
wall above its optimum temperature. A reduction in lamp life can result.
[0006] A proper balance must be maintained between minimizing the glow state duration and
electric field magnitude to maximize both lamp efficacy and lamp life. This balance
is difficult to achieve since a decrease in the amount of energy supplied to the electrode
will prolong the glow state duration while an increase in the amount of energy supplied
to the electrode will shorten lamp life through an increase in sputtering.
[0007] As compared to the non-thermionic glow phase, during the thermionic arc phase, the
lamp voltage is reduced and lamp current is increased. During the thermionic arc phase,
residual sputtering can still occur. After applying a sufficiently high current to
heat the electrode during the thermionic arc phase, the current is reduced and thereafter
the lamp is operated under steady state conditions.
[0008] Accordingly, it is desirable to provide an improved electronic ballast in which the
variation in component tolerances can be decreased. The ballast should also provide
a proper balance between minimizing the glow state duration and electric field magnitude
to maximize both lamp efficacy and lamp life.
[0009] Generally speaking, in accordance with a first aspect of the invention, an electronic
ballast for starting a high intensity discharge lamp having breakdown, non-thermionic
glow and thermionic arc phases prior to entering a steady state operating mode includes
a resonant ignitor, an exciter and a controller. The resonant ignitor applies a lamp
voltage across the lamp. The exciter is responsive to a varying control signal for
exciting the resonant ignitor at a varying operating frequency. The controller produces
the varying control signal based on the phase of the lamp. Prior to breakdown, the
controller adjusts the control signal so as to decrease the operating frequency so
as to sweep toward a lamp voltage equal to a predetermined ignition voltage. Upon
entering the glow phase, the control signal is further adjusted for a predetermined
time duration within the glow phase. Upon entering the thermionic arc phase, the controller
further adjusts the control signal so as to increase lamp current.
[0010] The controller by adjusting the control signal so as to decrease the operating frequency
in sweeping toward a lamp voltage equal to a predetermined ignition voltage can limit
the maximum starting voltage so that electronic ballast components need not be designed
to withstand voltages as high as 4 to 5 KV. The control signal also can be adjusted
so as to provide a proper balance between minimizing the glow state duration and electric
field magnitude to maximize both lamp efficacy and lamp life. In the non-thermionic
glow phase, the frequency is adjusted so as to reduce the voltage to a level which
is known to produce minimal blackening of the lamp surface while being sufficiently
high to support continued non-thermionic emission. Of particular importance, the controller
during the thermionic arc (i.e.run-up) phase further adjusts the frequency so as to
increase the current, at a reduced voltage, so as to support the onset of thermionic
emission, yet continue to minimize lamp blackening. Thereafter, the frequency is adjusted
by the controller so as to reduce the current and increase the voltage to the steady
state conditions for the lamp.
[0011] In a feature of the invention, the controller maintains the control signal at a preset
operating frequency for a prefixed period of time during which the lamp voltage is
held at the predetermined ignition voltage. By applying the predetermined voltage
for, at most, a prefixed period of time, the stress on all components is reduced thereby
extending their lives. In another feature of the invention, the controller can maintain
the control signal at a preset operating frequency for a prefixed period of time during
the non-thermionic glow phase. In yet another feature of the invention, the controller
can maintain the control signal at a preset operating frequency for a prefixed period
of time during the thermionic arc phase.
[0012] Accordingly, it is an object of the invention to provide an improved electronic ballast
which improves lamp life and minimizes lamp blackening of an HID-lamp.
[0013] It is a further object of this invention to provide an improved electronic ballast
which reduces the cost and complexity of the components required to effect turn on
of the lamp and to avoid the premature failure of these components.
[0014] The invention accordingly comprises several steps in a relation of one or more of
such steps with respect to each of the others, and the device embodying features of
construction, a combination of elements and arrangement of parts which are adapted
to effect such steps, all is exemplified in the following detailed disclosure and
the scope of the invention will be indicated in the claims.
Figure 1 is a block diagram of an HID electronic ballast in accordance with the invention;
Figures 2a, 2b and 2c illustrate the transfer characteristics of a resonant ignitor
during each phase of lamp turn on; and
Figure 3 is a flow chart illustrating the methodology employed in accordance with
the invention.
[0015] Figure 1 shows a block diagram of a preferred embodiment of a high frequency electronic
ballast for a high intensity discharge (HID) lamp. Lamp 9 is ignited by a resonant
ignitor 3, which receives its excitation voltage and current from an exciter (e.g.
a bridge circuit) 2. Exciter 2 is triggered by a control signal 40 which controls
the generation of a high current signal 20 which excites resonant ignitor 3. The frequency
of signal 20 and a resultant voltage 30 applied to lamp 9 is determined by control
signal 40. High current signal 20 is in the form of a pulse train. Both exciter 2
and a detector 5 are powered by a power source 1. Control signal 40 is supplied by
a controller (e.g. a microcontroller) 4. Frequency of operation by the exciter 2 is
determined by control signal 40.
[0016] Figure 2a shows the transfer characteristics of resonant ignitor 3 in combination
with lamp 9 prior to lamp breakdown. Lamp 9, in the non-conducting state, is effectively
an open circuit, and resonant ignitor 3 has the transfer characteristics of a conventional
resonant circuit. At a resonant frequency f
O, the voltage will be at a peak voltage Vp. Also shown in Figure 2a is a voltage V
MIN which represents the minimum specified voltage required to drive a lamp of a given
type into breakdown provided that voltage is applied for a minimum specified time
duration (not shown).
[0017] Ideally, resonant ignitor 3 is designed so that voltage V
P is equal to voltage V
MIN, that is, so that control signal 40 excites resonant ignitor 3 at exactly resonant
frequency f
O whereby only the required minimum voltage V
MIN is applied. The value of the particular components within ignitor 3 will determine
the particular value of frequency f
O, as well as a width d of a waveform about resonant frequency fo and the magnitude
of peak voltage V
P.
[0018] As shown in figure 2a, exciting the circuit ofresonant ignitor 3 at a frequency of
f
X produces a voltage V
X. If the specific values of frequency f
O and width d are such that voltage V
X is below voltage V
MIN, as shown in figure 2a, the lamp may not ignite. Conventional resonant ignitors are
designed to assure that regardless of the variations in component values, frequency
f
O is relatively constrained, and voltage V
P is sufficiently large so as to include a wide range of excitation frequencies at
which voltage Vx can be at least equal to voltage V
MIN· Constraining the frequency f
O requires components with tight tolerances which increases the cost of the components.
The peak voltage V
P of the conventional resonant ignitor is designed to be at least about one and one
half to three times the required minimum voltage V
MIN to compensate for the deviation in breakdown characteristics among different lamps.
Conventional ignitors are designed to operate at the peak voltage Vp for the specified
minimum time duration in the event of excitation at about the resonant frequency fo.
That is, the conventional resonant ignitor is designed to operate at twice the required
voltage, introducing additional costs, as well as reduced component life.
[0019] In accordance with this invention, however, resonant ignitor 3 is operated at the
minimum required voltage level V
MIN thereby avoiding the aforementioned additional costs and additional component stress.
Controller 4 controls control signal 40 so as to initially excite the resonant ignitor
at a frequency f
H well above the nominal resonant frequency. Controller 4 thereafter reduces the excitation
frequency while monitoring through a detector 5 the voltage across lamp 9. By reducing
the frequency toward the nominal resonant frequency, the voltage across lamp 9 will
increase. When the measured voltage across lamp reaches V
MIN, at frequency f
i, controller 4 ceases reduction in frequency and continues to excite the lamp at this
frequency for the specified minimum duration.
[0020] Voltage V
MIN and the minimum duration are the specified values at which all lamps of a given type
are assured to ignite; some lamps will ignite at lower voltages and at shorter durations.
Detector circuit 5 detects the flow of current through lamp 9. When lamp 9 ignites,
current flows through lamp 9. Detector circuit 5 provides a current detected signal
supplied to controller 4 once current begins to flow through lamp 9. In response to
this signal, controller 4 effects the non-thermionic glow phase. Ignitor 3 operates
at a frequency which produces voltage V
MIN until lamp current flows or until the minimum non-thermionic time duration elapses,
whichever occurs sooner.
[0021] When lamp 9 does not enter breakdown there is no lamp current. The minimum time duration
for breakdown will elapse. To minimize thermal stress, controller 4 pauses for a few
moments to allow the components in the ballast to cool; thereafter, repeats the above
process, starting with an excitation frequency of f
H. When the frequency is swept down to f
L (i.e. below the resonant frequency fo) before the resonant circuit produces the required
minimum voltage V
MIN, controller 4 ceases the frequency sweep and restarts the above process. Typically,
when voltage V
MIN has not been achieved, there is no need to pause to allow the components to cool.
[0022] The above process is continually repeated until the flow of lamp current is detected,
at which point the lamp enters the non-thermionic glow phase. At breakdown, the impedance
of lamp 9 changes from a virtual open circuit to a conductor of lamp current, thereby
loading resonant ignitor 3.The transfer function associated with the glow phase is
shown in figure 2b. The transfer function includes a relatively flat section 230 at
low frequencies with a peak 235 about frequency f
O and thereafter falls off at a sloping section 237 for frequencies above frequency
fo.
[0023] Parts of the electrode are sprayed or sputtered during the glow phase. The loss of
material from the electrode will eventually lead to electrode failure, and the lost
material adheres to the inner surface of the lamp, causing a blackening which decreases
the amount of light emitted from the lamp. By design, or via life testing, an optimal
glow phase voltage can be determined so as to maximize the effective life of the electrode
and minimize the blackening of the lamp. In accordance with this invention, the frequency
of operation will be adjusted so as to provide this optimal voltage to the lamp during
the glow phase.
[0024] Depending upon the particular characteristics of the lamp, the optimal voltage can
be achieved in a variety of ways. If there is very little variability in the transfer
function for a given lamp type, specific frequency f
OPT associated with the optimal voltage V
OPT can be predefined in controller 4. Upon entering the glow phase, controller 4 will
adjust control signal 40 so as to have resonant ignitor 3 produce optimal frequency
f
OPT. When there is significant variability in the transfer function of lamps of the same
type, an iterative approach can be employed wherein controller 4 sets the operating
frequency to the nominal f
OPT, as discussed above. Thereafter, detector 5 compares the lamp voltage to voltage
V
OPT and adjusts the excitation frequency to achieve voltage V
OPT. As shown in Figure 2b, frequency f
OPT is in the sloping section (falloff) 237 of the curve. When the measured voltage is
below voltage V
OPT, the frequency of operation is reduced. When the measured voltage is above V
OPT, the frequency of operation is increased. As in the breakdown phase, a lower frequency
limit f
L is set, beyond which the frequency is not adjusted.
[0025] The voltage is maintained at the optimal level throughout the glow phase. In accordance
with the invention, when design or life tests indicate that the optimal voltage should
be increased or decreased during the glow period, controller 4 is programmed to effect
the tracking to this changing optimal level. In a first embodiment, the voltage is
maintained at the optimal glow phase voltage for a predetermined time period (e.g.
about one half second). In an alternative embodiment, in response to an increase in
lamp current sensed by detector 5 when entering the thermionic arc phase, controller
4 readjusts the frequency of exciter 2 operation. During the glow phase as well as
all subsequent phases, lamp current is continually monitored. When lamp current ceases,
the entire process is repeated starting with the breakdown phase.
[0026] When thermionic emissions begin, at the start of the thermionic arc phase, lamp current
increases and lamp voltage significantly decreases. Again, by design or life tests,
an optimal run-up current I
OPT can be determined such that lamp 9 is brought to full thermionic emission as quickly
as possible without causing excessive electrode degradation and blackening. In one
embodiment, a run-up current of approximately 1.25 times to about twice the steady
state current significantly reduces degradation and blackening. Following the thermionic
arc (i.e. run-up) phase, lamp current is reduced to its steady state operating current.
To effect this current control, either the current, voltage or impedance can be monitored
with the frequency adjusted accordingly. Correlation between current and voltage for
a particular lamp type during the thermionic phase is determined by design or test
such that the run-up phase voltage V
R which produces I
OPT can be predefined. A typical transfer function corresponding to the thermionic arc
phase is shown in figure 2c. In operation, controller 4 adjusts the control signal
40 so as to adjust the operating frequency until the lamp voltage is equal to voltage
V
R..
[0027] Often a broad range of voltages or currents is available within which optimal performance
can be achieved while minimizing degradation and blackening. Under these conditions
or when optimal performance is not required, a nominal resonant frequency f
R can be defined with controller 4 programmed to adjust the operating frequency to
this nominal frequency. Iterative frequncy adjustment in achieving optimal values
for current or voltage is therefore not necessarily required in either the glow or
arc phase.
[0028] Figure 3 is a flowchart illustrating the method for controlling the starting of a
lamp. To ignite lamp 9, the operating frequency f is set to f
H at a step 310. The loop formed by a series of steps 320-328 decreases the frequency
until the lamp voltage reaches voltage V
MIN. The lamp voltage is measured at step 320. At a step 328, the operating frequency
is decreased provided that the lamp voltage has not yet reached voltage V
MIN (as determined at a step 324) and the operating frequency is not at or below lower
frequncy limit f
L (as determined at a step 326). When the lower frequency limit has been reached, the
frequency is reset to f
H at step 310. Lamp current is monitored throughout this process including at a step
315. Once lamp 9 ignites, as determined at step 315, the process enters the glow phase
through a step 340.
[0029] When the lamp voltage has reached voltage V
MIN and lamp current is not yet flowing, a duration timer is started at a step 330. The
process loops through a series of steps 334-336 until either lamp current flows as
determined at step 334 or the minimum lamp ignition duration DMIN has elapsed as determined
at step 336. When the minimum lamp ignition duration DMIN has elapsed, the process
enters a cool down period at a step 312 and is then restarted at step 310. When at
step 334 it is detemined that lamp current is flowing, the process enters the glow
phase through step 340.
[0030] At step 340, the operating frequency is set to the nominal glow phase optimal frequency
f
OPT. An iterative process can be employed, similar to the loop at steps 320-328 so as
to adjust the frequency to achieve voltage V
OPT. In a preferred embodiment of the electronic ballast, the operating frequency is
held at the nominal frequency f
OPT for a duration DOPT via steps 344-346. Lamp current is monitored at step 346. When
lamp 9 is extinguished, as indicated by the cessation of current, and following a
cool down period at step 312 the process is restarted through step 310. After the
duration DOPT elapses, the process enters the run-up phase through step 350.
[0031] At step 350, the operating frequency is set to the nominal run-up phase frequency
f
R. An iterative process can be employed similar to the loop at steps 320-328 to adjust
the frequency to achieve voltage V
R or current I
OPT. In the preferred embodiment, the operating frequency is held at the nominal frequency
f
R for a duration DR through a series of steps 354-356. After the duration DR elapses,
the process returns to a step 395 to enter the steady state mode of operation.
[0032] As can now be readily appreciated, through precise control by controller 4 of the
exciter frequency throughout the three phases of ignition, a reduction can be had
in lamp degradation and blackening. In particular, lower voltage and power ballasts
components can be employed. A reduction in the stress on these components is also
achieved.
[0033] It will thus be seen that the objects set forth above and those made apparent from
the preceeding description are efficiently attained. Certain changes can be made to
the electronic ballast providing the above method and of the construction set forth
without departing from the spirit and scope of the invention, for example, controller
4 can comprise a voltage or current detector obviating the need for a discrete voltage
comparator or current detector 5. Controller 4 can provide excitation signal 20 directly
to resonant ignitor 3; and ignitor 3 can be made from voltage multipliers which need
not achieve high voltage signals for lamp ignition through resonance. Although the
invention has been described in terms of high ignition voltage lamps, the principles
embodied herein are not limited to lamps which require high ignition voltages.